liveBook · Manning

GSD

As explained in Vibe Engineering

GSD is a spec-driven AI-agent development workflow that externalizes project information and coordinates task-oriented agents.

GSD is a spec-driven AI-agent development workflow that treats context rot as a workflow problem. It externalizes project intelligence into persistent files and uses disposable agent sessions for individual tasks. section 3.5.3

Use its lifecycle to move from product decisions to verified implementation:

  • Discuss: clarify questions, surface assumptions, and persist decisions in CONTEXT.md before writing code. section 3.5.3
  • Plan: research the codebase and generate a detailed implementation plan divided into small, concrete tasks. section 3.5.3
  • Execute and Verify: continue the workflow through implementation and validation rather than relying on one persistent conversation. section 3.5.3

Behind a small command surface, GSD combines XML prompt formatting, subagent orchestration, and state management. It runs with Claude Code and supports OpenCode, Gemini CLI, Codex, Copilot, Cursor, and Antigravity. section 3.5.3

How Vibe Engineering covers it

How deep it goes
Defines a four-stage lifecycle and names prompt, orchestration, and state mechanisms without detailing their implementations.
Examples
Uses CONTEXT.md, codebase research, implementation plans, specs, research notes, and validation contracts as persistent artifacts.
What you can do after
Operate a Discuss–Plan–Execute–Verify workflow across supported AI coding agents.
How it connects
Connects context rot with meta-prompting, state management, subagent orchestration, and context externalization.

Related concepts

This explainer was generated by Manning's knowledge graph from the text of Vibe Engineering. The section links point to the parts of the book it draws on.